Most process-based crop simulation models require input of detailed soil information. The large-scale State Soil Geographic (STATSGO) Database is available for the continental U.S., but lacks details in some soil parameters, such as water retention, that are critical for crop model performance. On the other hand, a finer-scale soil database, the Soil Survey Geographic Database (SSURGO), provides most details on soil characteristics, but is not yet available for the entire country. The purposes of this study were to assess the effects of soil database spatial scales on crop yield simulations, and to evaluate potential improvements when additional remotely sensed vegetation parameters were incorporated into the model. A prominent crop model, EPIC, was modified and applied to corn yield simulations. Two key biophysical products of the Moderate Resolution Imaging Spectroradiometer (MODIS), the Fraction of Photosynthetically Active Radiation (FPAR) and crop Leaf Area Index (LAI), were used as model inputs. EPIC was run with cross combinations of different scales of soil data and with/without remotely sensed crop parameters. Ground truth corn LAI and yields were measured during the 2005 growing season within multiple corn fields in Sangamon County, IL. Results will be presented on comparisons between field observations and modeled yields as well as between model simulations using different combinations of input data.